Skip to main content
Own toolchain · 03

Task dashboard with voice input

A voice note is enough: the bot transcribes, classifies, creates tasks and surfaces reminders on time.

voice instead of typingclassified tasks at no extra effort
Telegram voice message on a smartphone being transcribed, classified and stored as a task list on EU server infrastructure

Pain

Typing tasks costs time you don't have between meetings. Notes on paper get lost; voice memos in the phone are never replayed. Net effect: tasks never make it into the backlog and slip out of mind.

Solution

  • Receive voice notes via a Telegram bot, the most natural channel for input on the go.
  • Transcribe with Whisper, segment the result into individual tasks, classify into categories (work, calendar, finance, personal).
  • Create each task in the backlog with a due date when one is detectable.
  • Push reminders back via the same Telegram channel, driven by cron jobs, not a second tool.

How it works

  1. 01
    Receive voice note
    Telegram bot reacts to voice notes in the personal chat.
  2. 02
    Whisper transcript
    Audio is transcribed locally with self-hosted Whisper, no external API call.
  3. 03
    Task classification
    A self-hosted open-source LLM splits the text into individual tasks and assigns category and priority.
  4. 04
    Backlog entry
    Each task lands in the dashboard with a source link to the original voice note.
  5. 05
    Cron reminders
    Recurring jobs check due dates and push reminders back into Telegram.

Outcome

Voice notes replace typing on the go. Tasks land classified directly in the backlog instead of leaking onto paper.

Why this stack

Whisper (self-hosted)
Solid transcription in German and English, fully on our own hardware, no usage-based API cost.
Telegram bot
Low entry barrier, push notifications, mobile-first.
n8n
Cron jobs and reminder logic in one place instead of scattered scripts.
More about our AI & Automation services

Related use cases

WebhookPOST /lead{ }Validate✦ClaudeclassifyCRMappend: rowGGeminireplyGmailsend: replysuccess{ }Error: data{ }Error: CRMTelegram alerterror
Own toolchain · 01

n8n workflows with LLM classification

Webhook → validate data → Claude classifies (heat score, intent) → save to CRM → Gemini drafts a personalized reply → Gmail. Plus an error branch with a Telegram alert.

one workflowper integration instead of a separate service
n8nClaudeGeminiWebhooks
Learn more→
CASHFLOW · 12-month forecast
“Tax lever on the Q4 spike?”
→ §7g investment deduction
→ Special depreciation, equipment
Own toolchain · 02

Personal finance dashboard with LLM

Own financial data as the basis. LLM analyses on demand: forecasting, tax-optimization hints on profit spikes, 12-month cashflow preview.

12-month forecaston demand instead of quarterly meeting
Claude APIPythonPower BI
Learn more→